鄂西贫困县耕地利用转型空间分异及其影响因素

耕地利用转型对推动乡村转型发展,促进乡村社会经济水平的提高具有重要意义。为准确分析鄂西贫困县耕地利用转型的空间分异特征及其影响因素,该文从空间形态和功能形态2方面测度了耕地利用转型状况,利用空间自相关模型分析了耕地利用转型的空间分异特征,并利用径向基神经网络模型从人口、经济和产业结构3方面分析了耕地利用转型空间分异的影响因素效用。结果表明:从全局空间自相关来看,鄂西16贫困县2002-2013年耕地空间转型具有显著正相关效应,其中在2002-2005年正相关效应最强;而耕地功能形态转型呈现出逐渐减弱的空间正相关效应。从局部空间自相关来看,鄂西16贫困县耕地空间形态转型的低-低区分布主要在鄂西北地区,高-高区和低-高区主要分布在鄂西南地区;耕地功能转型的局部空间自相关性较低,空间异质性不明显,仅来凤县处于高-高区。在影响因素效用方面,人均固定资产投资和城市化率对耕地空间形态转型的影响效用最大,人口密度和第一产业占比影响效用最低;人口密度和人均固定资产投资对耕地功能形态转型的影响效用最大,城市化率、人均GDP、第一产业占比等影响因素效用较低。在影响因素的调控过程中,因素的影响效用与耕地利用转型调控程度呈正相关,因素的影响效用越高,对耕地利用转型调控力度越大。

Spatial difference and its influence factors of cultivated land transition of poverty counties in west of Hubei

With the rapid development of industrialization and urbanization, many problems around cultivated land such as illegal occupation and non-agriculture use come into being, leading to the contradiction between human beings and cultivated land, which finally promotes the cultivated land use transition.It is of great significance to carry out the research on the transition of cultivated land use in the rural areas for the transition of rural areas and social-economic development.In order to analyze the spatial difference and its influence factors of cultivated land transition of poverty counties in west of Hubei Province, this paper measures the transition of cultivated land use from 2 aspects: spatial form and function form.It analyzes the spatial distribution characteristics of the spatial autocorrelation model, and discusses the effect of the factors from 3 perspectives: population, economy and industrial structure by the radial basis function neural networks (RBFNN) model.The results show that: from the perspective of global spatial autocorrelation, the spatial transition of cultivated land in the 16 poverty counties in the west of Hubei Province from 2002 to 2013 has a clear positive correlation effect, which is the strongest from 2002 to 2005, and there is a certain spatial clustering.Also the function of cultivated land transition has a clear positive correlation effect.However, the influence is gradually weakened.From the perspective of local spatial autocorrelation, the low-low regions of cultivated land are mainly distributed in the northwest of Hubei, and high-high and low-high regions are distributed in the southwest of Hubei.The local spatial correlation of function transition of cultivated land is in a lower situation.The spatial heterogeneity is not obvious, and only Laifeng is in high-high areas.In terms of the influence factors of spatial difference of cultivated land use transition, the investment and urbanization rate have a greater impact, while the effects of the population density and the first industry are less.For the utility of functional transition, the influences of population density and per capita fixed-asset investment have the strongest effect, while the effects of urbanization rate, per capita GDP (gross domestic product) and the first industry are comparatively lower.It can be seen that the impacts of different factors vary a lot for different objectives, which reflects that influence factors have different driving forces for space transition and function transition of cultivated land.In the process of the transition and development of cultivated land use, there is a positive correlation between influence factors and their driving forces.Therefore, it is necessary to adopt different strategies according to the driving force of different factors, in order to achieve the optimal solution of cultivated land use transition.Above all, the spatial autocorrelation model can be used in analyzing the spatial autocorrelation effectively for cultivated land use transition in poverty counties in western Hubei.The RBFNN model can measure the influence factors'' driving forces in the target of space transition and function transition of cultivated land effectively, which takes into account both original effect of influence factors and internal structure of cultivated land use transition affected by influence factors, carries out more comprehensive analysis, and provides a scientific reference for relevant research.